everyvoice segment files are being produce in the "audio" directory VS "." ( CWD) & why keep the resampled audio file once completed?

Just noticing a bug I think in segment.

Output files are being produced into the `AUDIO_PATH` folder . I was expecting the TextGrid file to be produce in the same directory where the program was executed .

Also, I think we should delete the converted audio file that was created in this example "`wavs/LJ002-0011-16000.wav`" once the task is completed by default.

```
echo "For two more years these high figures were steadily maintained, and in eighteen oh three the total rose to seven hundred ten." > foo.txt
(EveryVoice) [U20-GPSC7]:$ everyvoice segment foo.txt wavs/LJ002-0011.wav
loading model...
resampling audio from 22050 to 16000
processing text
performing alignment
creating textgrid
writing file to wavs/LJ002-0011-16000.TextGrid

(EveryVoice) [U20-GPSC7]:$ ll wavs/LJ002-0011-*
-rw-r----- 1 tes001 nrc_ict 300K Apr 22 15:46 wavs/LJ002-0011-16000.wav
-rw-r----- 1 tes001 nrc_ict 46K Apr 22 15:47 wavs/LJ002-0011-16000.TextGrid

```

Start by tracing the `everyvoice segment` command and the code that uses `AUDIO_PATH` during the sample invocation with `foo.txt` and `wavs/LJ002-0011.wav`. Check how the TextGrid destination and `wavs/LJ002-0011-16000.wav` lifecycle are selected. Done means the TextGrid is written to the intended working directory and the resampled file is removed by default after completion.

Written by the indexing model from the issue text.